With cluster auto-scaling, the actual load of your worker-nodes will be monitored actively. By adding and removing worker-nodes from the cluster, it ensures that enough resources are available to keep your application healthy and responsive. In contrast, it removes worker-nodes from the AKS cluster, to optimize resource utilization and be as cost-effective as possible Reference: https://docs.microsoft.com/en-us/azure/aks/cluster-autoscaler https://thorsten-hans.com/aks-cluster-auto-scaler-inside-out